How can banks defend against AI-driven fraud?

Almost half of Asia-Pacific banks lose more than $10m to fraud yearly.

Banks face growing risks from artificial intelligence (AI)-enabled fraud as criminals use the technology to create fake identities, forged documents, and deepfake videos, whilst many lenders report rising fraud losses and attacks.

“What is happening is that the methods they use to forge documents are perhaps more sophisticated, but also the methods of fighting fraud are becoming more sophisticated,” Konstantin Poptodorov, director of fraud and identity at LexisNexis Risk Solutions, Inc., told Asian Banking & Finance via Zoom. “So we’re kind of learning on the same level.”

Poptodorov said AI is now involved in almost every fraud type his team investigates. He cited fake identities, fake documentation, deepfake videos, and cases where people sell their identities for fraudulent activity.

A January 2026 survey by BioCatch Ltd. found that 86% of banking leaders in Singapore, Indonesia, Thailand, India, and Australia view AI agents as the next major vulnerability that fraudsters could exploit within the next year.

Among 340 banking leaders surveyed, 49% said they lose more than $10m to fraud annually, whilst 22% reported losses exceeding $25m.

BioCatch also found that 81% of respondents said fraud attempts are increasing, whilst 78% said fraud losses are rising.

The financial impact is already evident. Commonwealth Bank of Australia reported about $717m (A$1b) in potential fraudulent home loans to authorities in March. Some of the loans involved AI-generated documents, according to media reports.

Poptodorov said Australia's banking sector has not been overwhelmed by AI-enabled fraud, suggesting anti-fraud controls are helping lenders keep pace with increasingly sophisticated attacks.

Banks are also deploying AI to detect suspicious activity, creating an arms race between criminals and fraud-prevention teams, he added.

Questions to ponder:

• Are banks investing in fraud prevention as quickly as they are adopting AI tools?
• Will AI widen the gap between banks with advanced fraud controls and those without them?
• How should regulators respond as AI-enabled fraud becomes more sophisticated?